feat(providers): add ordered fallback sequence for primary model - #1761
Closed
uzaylisak wants to merge 2 commits into
Closed
feat(providers): add ordered fallback sequence for primary model#1761uzaylisak wants to merge 2 commits into
uzaylisak wants to merge 2 commits into
Conversation
teknium1
added a commit
that referenced
this pull request
Mar 29, 2026
Extends the single fallback_model mechanism into an ordered chain.
When the primary model fails, Hermes tries each fallback provider in
sequence until one succeeds or the chain is exhausted.
Config format (new):
fallback_providers:
- provider: openrouter
model: anthropic/claude-sonnet-4
- provider: openai
model: gpt-4o
Legacy single-dict fallback_model format still works unchanged.
Key fix vs original PR: the call sites in the retry loop now use
_fallback_index < len(_fallback_chain) instead of the old one-shot
_fallback_activated guard, so the chain actually advances through
all configured providers.
Changes:
- run_agent.py: _fallback_chain list + _fallback_index replaces
one-shot _fallback_model; _try_activate_fallback() advances
through chain; failed provider resolution skips to next entry;
call sites updated to allow chain advancement
- cli.py: reads fallback_providers with legacy fallback_model compat
- gateway/run.py: same
- hermes_cli/config.py: fallback_providers: [] in DEFAULT_CONFIG
- tests: 12 new chain tests + 6 existing test fixtures updated
Co-authored-by: uzaylisak <uzaylisak@users.noreply.github.com>
teknium1
added a commit
that referenced
this pull request
Mar 29, 2026
…3813) Extends the single fallback_model mechanism into an ordered chain. When the primary model fails, Hermes tries each fallback provider in sequence until one succeeds or the chain is exhausted. Config format (new): fallback_providers: - provider: openrouter model: anthropic/claude-sonnet-4 - provider: openai model: gpt-4o Legacy single-dict fallback_model format still works unchanged. Key fix vs original PR: the call sites in the retry loop now use _fallback_index < len(_fallback_chain) instead of the old one-shot _fallback_activated guard, so the chain actually advances through all configured providers. Changes: - run_agent.py: _fallback_chain list + _fallback_index replaces one-shot _fallback_model; _try_activate_fallback() advances through chain; failed provider resolution skips to next entry; call sites updated to allow chain advancement - cli.py: reads fallback_providers with legacy fallback_model compat - gateway/run.py: same - hermes_cli/config.py: fallback_providers: [] in DEFAULT_CONFIG - tests: 12 new chain tests + 6 existing test fixtures updated Co-authored-by: uzaylisak <uzaylisak@users.noreply.github.com>
Contributor
|
Merged via #3813. Your fallback chain concept was salvaged onto current main with the call-site bug fixed (chain now actually advances through all providers). Thanks for the contribution! Credit preserved via Co-authored-by. |
angelburgosrosado
pushed a commit
to angelburgosrosado/hermes-agent
that referenced
this pull request
Apr 27, 2026
…earch#1761) (NousResearch#3813) Extends the single fallback_model mechanism into an ordered chain. When the primary model fails, Hermes tries each fallback provider in sequence until one succeeds or the chain is exhausted. Config format (new): fallback_providers: - provider: openrouter model: anthropic/claude-sonnet-4 - provider: openai model: gpt-4o Legacy single-dict fallback_model format still works unchanged. Key fix vs original PR: the call sites in the retry loop now use _fallback_index < len(_fallback_chain) instead of the old one-shot _fallback_activated guard, so the chain actually advances through all configured providers. Changes: - run_agent.py: _fallback_chain list + _fallback_index replaces one-shot _fallback_model; _try_activate_fallback() advances through chain; failed provider resolution skips to next entry; call sites updated to allow chain advancement - cli.py: reads fallback_providers with legacy fallback_model compat - gateway/run.py: same - hermes_cli/config.py: fallback_providers: [] in DEFAULT_CONFIG - tests: 12 new chain tests + 6 existing test fixtures updated Co-authored-by: uzaylisak <uzaylisak@users.noreply.github.com>
02356abc
pushed a commit
to 02356abc/hermes-agent
that referenced
this pull request
May 14, 2026
…earch#1761) (NousResearch#3813) Extends the single fallback_model mechanism into an ordered chain. When the primary model fails, Hermes tries each fallback provider in sequence until one succeeds or the chain is exhausted. Config format (new): fallback_providers: - provider: openrouter model: anthropic/claude-sonnet-4 - provider: openai model: gpt-4o Legacy single-dict fallback_model format still works unchanged. Key fix vs original PR: the call sites in the retry loop now use _fallback_index < len(_fallback_chain) instead of the old one-shot _fallback_activated guard, so the chain actually advances through all configured providers. Changes: - run_agent.py: _fallback_chain list + _fallback_index replaces one-shot _fallback_model; _try_activate_fallback() advances through chain; failed provider resolution skips to next entry; call sites updated to allow chain advancement - cli.py: reads fallback_providers with legacy fallback_model compat - gateway/run.py: same - hermes_cli/config.py: fallback_providers: [] in DEFAULT_CONFIG - tests: 12 new chain tests + 6 existing test fixtures updated Co-authored-by: uzaylisak <uzaylisak@users.noreply.github.com>
gweeteve
pushed a commit
to gweeteve/hermes-agent
that referenced
this pull request
Jun 2, 2026
…earch#1761) (NousResearch#3813) Extends the single fallback_model mechanism into an ordered chain. When the primary model fails, Hermes tries each fallback provider in sequence until one succeeds or the chain is exhausted. Config format (new): fallback_providers: - provider: openrouter model: anthropic/claude-sonnet-4 - provider: openai model: gpt-4o Legacy single-dict fallback_model format still works unchanged. Key fix vs original PR: the call sites in the retry loop now use _fallback_index < len(_fallback_chain) instead of the old one-shot _fallback_activated guard, so the chain actually advances through all configured providers. Changes: - run_agent.py: _fallback_chain list + _fallback_index replaces one-shot _fallback_model; _try_activate_fallback() advances through chain; failed provider resolution skips to next entry; call sites updated to allow chain advancement - cli.py: reads fallback_providers with legacy fallback_model compat - gateway/run.py: same - hermes_cli/config.py: fallback_providers: [] in DEFAULT_CONFIG - tests: 12 new chain tests + 6 existing test fixtures updated Co-authored-by: uzaylisak <uzaylisak@users.noreply.github.com>
waefrebeorn
pushed a commit
to waefrebeorn/slermes
that referenced
this pull request
Jul 2, 2026
…earch#1761) (NousResearch#3813) Extends the single fallback_model mechanism into an ordered chain. When the primary model fails, Hermes tries each fallback provider in sequence until one succeeds or the chain is exhausted. Config format (new): fallback_providers: - provider: openrouter model: anthropic/claude-sonnet-4 - provider: openai model: gpt-4o Legacy single-dict fallback_model format still works unchanged. Key fix vs original PR: the call sites in the retry loop now use _fallback_index < len(_fallback_chain) instead of the old one-shot _fallback_activated guard, so the chain actually advances through all configured providers. Changes: - run_agent.py: _fallback_chain list + _fallback_index replaces one-shot _fallback_model; _try_activate_fallback() advances through chain; failed provider resolution skips to next entry; call sites updated to allow chain advancement - cli.py: reads fallback_providers with legacy fallback_model compat - gateway/run.py: same - hermes_cli/config.py: fallback_providers: [] in DEFAULT_CONFIG - tests: 12 new chain tests + 6 existing test fixtures updated Co-authored-by: uzaylisak <uzaylisak@users.noreply.github.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #1734
What
Adds support for an ordered
fallback_providerslist in config.yaml. When the primary model fails (rate limit, outage, connection error), Hermes automatically tries each provider in sequence until one succeeds.Changes
run_agent.py:_fallback_chainlist replaces single_fallback_modeldict;_try_activate_fallback()now advances through the chainhermes_cli/config.py: Addedfallback_providers: []to DEFAULT_CONFIGcli.py: Readsfallback_providerslist (falls back to legacyfallback_modelfor compat)gateway/run.py: Same — readsfallback_providerslist with legacy fallbacktests/test_provider_fallback.py: Unit tests for chain init, backwards compat, sequence exhaustionwebsite/docs/user-guide/features/fallback-providers.md: Updated docs with list format and new examplesUsage
Backward Compatibility
Old single-dict
fallback_modelformat still works unchanged.